Introduction

In this BLS Data Quest, we explore the new Job Openings and Labor Turnover Survey (JOLTS) experimental state estimates found here. I chose to tackle this problem by building an animated scatterplot of the Job Openings Rate vs the Layoff and Discharge Rate. You can see how the these numbers relate to each other over time. Hit Play to begin the animation. You can also use the slider to move to specific dates.

The black line in the plot below is the line y=x, that is where job openings equals the number of layoffs/discharges. Points above this line are laying off more than they are growing. A good example of this is if you move the slider to the year 2009 to watch the Great Recession play out.


Another observation to make is that the outlier Alaska (AK) dashes wildly back and forth across the scatterplot. Why is this? The series happens to be very seasonal as you can see in the time-series plots below.



You can find a few other states that exhibit this seasonality, including Montana (MT). To fix this, we use the Census Bureau’s X-13-ARIMA-SEATS software to seasonally adjust the series via the R seasonal package. This way the seasonality will not obscure true economic movements that would be of interest to an analyst (A fair warning is that one should be more careful than I have seasonally adjusting, this is for the sake of quick analysis). Alaska remains an outlier which I hypthesize to be due to high turnover in employment. We recreate the animation below.


Explore the animations and see if you can ask some questions of your own.